Economic Forecast Based On Time Series

In recent years, plenty of practical issues require modeling analysis and accurateforecast of Times Series data. The R-based General ARIMA Model andRegression-Time Series Model puts up modeling theoretical basis, specific modelingsteps, examining methods, screening principles and forecast.In addition, with these theories, this thesis demonstrates R-based modelingmethods for China’s export trade volume between January2001and May2012and acity’s tax revenue between1994and2011while providing forecast value, interval andaccuracy for these models.Results indicate that the General ARIMA Model and Regression-Time SeriesModel have relatively high forecast accuracy. However, forecast accuracy decreasesas forecast step increases. Therefore, these two models are more suitable forshort-term forecast with historical data.